In conventional graphics processing architectures, a software (SW) driver running on a central processing unit (CPU) may submit workloads to a graphics processing unit (GPU) for high performance execution. During workload execution by the GPU, “synchronization” conditions may occur as particular execution points are reached or completed, wherein such execution points may be referred to as “synchronization points”. The ability of the software driver to use post-synchronization operations to track the progress made by the GPU may be limited to simple memory writes. As a result, sub-optimal performance and/or power consumption may be experienced.